'HomeAssistantCore.update' blocked from execution, system is not healthy

I am using the supervised method in Rpi4 and SSD. If I want to switch to HA OS, what will happen with the following items:

  1. My python script that is running from root access and will communicate with a CAN-bus controller.
  2. HACS and custom components
  3. SSD drive instead of SD card

Are they still be working in HA OS?

Baaad decisions are being made here. You will alienate tons of users like myself.

You guys are operating on a VERY slippery slope. Im not a super advanced user but I can find my way around eventually. Yeah, I found this thread, but others may not. And yes, I run a NUC with Debian and Portainer and run everything in my shop and solar system with home assistant. It’s not a hospital or airport. If some lights don’t turn on, no big deal.

Disabling health checks is a REALLY bad idea. I used that last week to finally fix MQTT changes you all depricated.

Home assistant used to be fun. Roadblocks like this turn it into a chore and a big PITA. Why do you want to make users like me curse your platform? Microsoft already has that in spades.

Rethink this guys. Seriously.

Jeff
Home assistant user from the very beginning

6 Likes

I don’t think there will be any turning back from this, just like when it happened with the “unsupported” state. But at least then we didn’t have to stop portainer, restart supervisor, update and then restart portainer, in order to do a simple update. It will only get harder from now on.
Just to clarify: By just having any container installed alongside HA makes the instance “unsupported”! As soon as portainer starts, the instance becomes “unhealthy” too!

Yes, going forward in this direction I think very soon the userbase number and rate of new users will start to decline.

  • Newcomers and people interested in trying it out will be deterred by the fact that they are more or less forced to get a dedicated machine for HA, unless they are willing to spend countless hours trying to get around the increasing number of “show stoppers”.

  • Existing users will grow increasingly frustrated and might consider looking elsewhere

The attitude in itself is detering, it’s a community driven project and still there are design decisions being made that will shut out more and more users, users who’s been giving and receiving support for years will find themselves in a whole new situation where they instead will be struggling just to keep their HA installations running.

Not to late to go back though, continue to warn people that they are running an unsupported instance but let the updates go through.
Make the Unhealthy information more visible, create pop-ups, have red flags everywhere, right now it’s really counterintuitive finding that info as @Tryfos pointed out.

6 Likes

Thank you everybody, solved stopping portainer, update home assistant and restarting portainer.

stop portainer, restart supervisor, update and then restart portainer

did that with the first update, but then I just removed portainer and… started it again, with different container name. problem solved, system healthy, next update worked normally with running portainer :slight_smile:

1 Like

Does this actually work consistently? How long did you wait after you restarted HA (with Portainer’s new name I mean)? In my experience it needs some time before the system is marked as ‘unhealthy’, so maybe you just updated within this time frame and it was then marked as unhealthy too.

To what other extremes are we going to go, in order to circumvent these unneeded and unjustified changes…I already feel awkward discussing this kind of solutions here, since the thread is being actively monitored by the developers, who I fear will immediately run to patch any loopholes… We’re slowly descending into ‘how to avoid an online Microsoft account for Win11 installation’ territory here and it kinda sucks to be honest.

when I was processing those changes - didn’t wait any special amount of time, but doing this while I was at work gave me some “natural breaks” between this or that command.

as for now - two machines working for almost a week without issues. “unsupported”? yes [ubuntu server + few own dockers], “unhealthy”? nope :slight_smile:

I totally agree. first I was going to yell how I feel about this changes, but I realized that it won’t change a thing… I do plan going to VM tho, to secure myself from any “patches” and “it’s for our good” changes in the HA, that will prevent me from using my own hardware the way I like, but right now went to quick’n’dirty scenario :wink:

Yeah, I’m afraid that supervised installs are not to last and/or to be supported for a long time and I’m starting to look elsewhere, especially after the last change, which has made obvious to me that the project is heading a direction I don’t like. But that’s my opinion, I know that developers are not particularly interested in our opinions.

Btw, why are you planning to go for a VM and not a docker install? That would be option no 2 for me after the supervised install.

knowing that for the last few years all I bought for my home was with HA in mind, I don’t even want to think of trying other solutions… well, not again for sure :wink:

I’m naive, but I hope that sometime soon someone will take our “whining” for consideration… their goals are somewhat good, but execution not always fits the community IMHO.

honestly? seemed easier to maintain “fully working” VM installation - I mean addon management is easier with supervisor [or at least that’s how I see it. maybe I’m wrong and “manual” management and connection via docker isn’t that bad]. I’m not an expert in VM or docker, so tried to find pros and cons for both :wink:

1 Like

I’m a supervisor developer. I can assure you, we will not. We’re not the police or your parents.

The message has been conveyed. Many people removed portainer. Many did not. We’re aware the ones that did not are going to do what they want. Which is completely fine, do what you want, it’s your system.

There’s really only one reason that would change. If users begin submitting issues and hide the fact they’re using portainer and that turns out to be the root cause. Then we waste a lot of time debugging something which isn’t supported. Anytime that happens a check is added to prevent future wastes of time.

If you still want to run portainer after this, go for it. Idk if you noticed but I’m the one who shared how to turn off the healthy check around these threads. Just respect our time and don’t submit issues from unsupported systems that are being disguised as supported.

6 Likes

I renamed the container as well, but supervisor still sees the image of portainer, and that’s my problem
have you renamed just the container or as well the image?

Oh, I see. So it’s just a matter of time then (there always going to be one who does the boo-boo and then everyone take the fall for them). My arguments above have essentially not been answered and I politely stand by them. With this way of thinking, what’s to stop you from completely obstructing HA from even starting, when another container is “detected”? Absolutely nothing and the gradual descent to something like that is rather obvious. Where/when is this going to stop, if not at a complete “prohibition” of supervised installs?

Anyway, I still cannot understand for the life of me why you don’t simply add a switch or an option along with a big red flag. Judging by the above posts the overwhelming majority of the community is against this particular change. Instead, the “healthy” state is effectively hidden behind 5 clicks, an unrelated tab (Repairs?) and 3 dots, in order to reach it. This is at least counter-intuitive.
At least think about it.

Finally, why are you still proposing to completely shut off the health check and potentially create more problems for the system when some other problem comes along? We lose an important check. Since a better solution has come up (just renaming the container), I think that this should be the recommended one.

assuming that you’ve pulled the image with docker pull portainer/portainer-ce:latest, just after tagging and running portainer with other name, try this: docker rmi portainer/portainer-ce [or change names accordingly if your image is different].
worked for me.

what I did was to delete the container, and created a new container with a different name but with the same image
if I run your command I get following error message

Error response from daemon: conflict: unable to remove repository reference “portainer/portainer-ce” (must force) - container 11f804a25e25 is using its referenced image 033d989d812e

didn’t saw the exact WTH, so I made my own. feel free to vote here

I guess the important question here is, “does HA check only container names or image names too, and if yes to which extent”. If I were to rename my container to, say, P0rtainer (with an 0 instead of an o), would that be sufficient to bypass the check?

UNHEALTHY_IMAGES = [
    "watchtower",
    "ouroboros",
    "portainer",
] 

1 Like

I’m seeing where you’re getting at and I was thinking of asking it: Ok then, could someone give some advice/guidance, how to modify the code and remove these completely without breaking the system? I realise that this would have to be done after every update, but I guess it could be a temporary solution.

PS: My container is named portainer-ce and it also gets banned by the HA “police” (just kidding here :grinning: ).

Already did! :grinning: